Sie sind auf Seite 1von 3

Proceedings of PAC07, Albuquerque, New Mexico, USA WEOAKI02

OBSERVATIONS OF UNDERDENSE PLASMA LENS FOCUSING OF


RELATIVISTIC ELECTRON BEAMS∗

M.C. Thompson† ‡ , UCLA, Los Angeles, CA, 90095 and LLNL, Livermore, CA, 94550, USA
H. Badakov, J.B. Rosenzweig, G. Travish, UCLA, Los Angeles, CA, 90095, USA
R. Fliller, G. M. Kazakevich, P. Piot, J. Santucci, FNAL, Batavia, Illinois, 60510, USA
J. Li, R. Tikhoplav§ , University of Rochester, Rochester, New York 14627, USA

Abstract case the strong electric field created by the space charge
of the electron beam ejects the plasma electrons from the
Focusing of a 15 MeV, 19 nC electron bunch by an un-
beam region entirely, leaving a uniform ion column. It can
derdense plasma lens operated just beyond the threshold
be shown that the radial-focusing electric-force of this ion
of the underdense condition has been demonstrated in ex-
column is linear and given by F radial = −2πnp e2 r, where
periments at the Fermilab NICADD Photoinjector Labora-
e is the electron charge. An underdense plasma lens is,
tory (FNPL). The strong 1.9 cm focal-length plasma-lens
therefore, theoretically free of spherical aberration in the
focused both transverse directions simultaneously and re-
limit that the ion column is as wide as the beam and the
duced the minimum area of the beam spot by a factor of 23.
ions are immobile. The head of the beam is not focused in
Analysis of the beam-envelope evolution observed near the
an underdense plasma lens because of the finite response
beam waist shows that the spherical aberrations of this un-
time of the plasma electron density distribution. Focusing
derdense lens are lower than those of an overdense plasma
by overdense plasma lenses of up to 4 MT/m strength has
lens, as predicted by theory. Correlations between the beam
been demonstrated in previous experiments [5 - 9]. Lens-
charge and the properties of the beam focus corroborate this
configuration focusing in the low-aberration underdense-
conclusion.
plasma-lens regime, however, has not been previously ob-
served.
INTRODUCTION
Plasma lenses are of great interest to the relativistic EXPERIMENTAL METHODOLOGY
beam physics community because they can provide radially
symmetric focusing gradients equivalent to a quadrupole In the experiment, a 14.8 MeV electron beam is focused
lens gradient of the order 1 MT/m, which exceeds the by a plasma lens of approximately gaussian profile in z
strength of conventional devices by many orders of mag- with FWHM of 19.3 mm and peak density n p,peak = 4.9
nitude [1, 2]. Additionally, it has been shown theoret- x 1012 cm−3 . This plasma lens is uniform in r on the
ically that adiabatic plasma lenses [3] can overcome the length scale of the beam, has an average focusing strength
synchrotron radiation-induced limit on final focus spot size K = 2561 m −2 over an effective length of l = 20.5 mm,
[4]. Consequently, there is great interest in using plasma and a focal length (f = 1/Kl) of 1.9 cm. In magnetic
lenses to achieve the small spot sizes and high luminosity quadrupole units, the lens has an average focusing gradi-
necessary at the interaction point of future e + e− colliders ent of 126 T/m, which is about 40 times stronger than the
such as the proposed International Linear Collider (ILC). magnetic qudrupole fields used to transport the beam to our
We recently made the first observation of plasma lens fo- experiment.
cusing of a relativistic electron beam in the low-aberration The plasma lens was created using a direct-current
underdense-regime: n b  np /2, where nb and np are the
beam and plasma densities, respectively.
Plasma lenses operate in either the overdense (n b << Bulk Plasma Flow
np ) or underdense regime. In the overdense case the plasma
Conical Translatable Mask
electrons spatially configure so that the plasma ions cancel
Barrier
the beam space charge and thus allow the beam to focus
r
under its magnetic self-forces. These self-forces are not
linear in distance from the beam axis (r), or uniform in dis- z
tance along the beam axis (z), which leads to significant e-
aberrations in the overdense case [2]. In the underdense Collimating Lens
Movable Plasma
Column
∗ Work supported by U.S. Dept. of Energy Contracts No. DE-FG03- Light to Camera
Fixed OTR Screen
92ER40693 and W-7405-ENG-48 (Streak Camera)
† dr.mcthompson@gmail.com
‡ Now at TAE Inc., Irvine, CA
§ Now at UCLA Figure 1: Schematic of the plasma lens apparatus.
03 Linear Colliders, Lepton Accelerators and New Acceleration Techniques A14 Advanced Concepts
1-4244-0917-9/07/$25.00 2007
c IEEE 1907
WEOAKI02 Proceedings of PAC07, Albuquerque, New Mexico, USA

electrical-discharge transversely offset from the beam or- plasma lens.


bit as illustrated in Fig. 1. A bulk plasma column of ap-
proximate width 5 cm FWHM is produced using our pulsed UNDERDENSE PLASMA FOCUSING
argon-discharge plasma-source [10] and allowed to diffuse
toward the beam path under the weak confinement of a 53 In order to understand the spherical aberrations of our
G solenoidal magnetic field aligned with the plasma flow. plasma lens we made a series of round beam focusing
A conical stainless steel barrier suppresses diffusion into measurements with different spacings between the plasma
the interaction area and a thin slice of the static plasma col- lens and OTR screen. These experiments used a 14.8
umn is selected using a translatable mask with a 1.25 cm MeV bi-gaussian electron beam with initial dimensions σ x
wide slit. This arrangement allows the distance between = 692 ± 54 μm, σy = 656 ± 51 μm, Q = 18.8 ± 2.0 nC, and
the plasma lens column and the fixed optical transition ra- σt = 22 ± 3 ps. Consequently the initial peak beam den-
diation (OTR) diagnostic screen to be varied without al- sity was nb,peak = 2.5 ± 0.5 x 1012 cm−3 . Uncertainties
tering the OTR collection optics. An in-vacuum, 160 mm are dominated by shot-to-shot variation of the beam. Peak
focal length, 2 inch diameter lens provides quasi-parallel density of the plasma lens is n p,peak  4.9 x 1012 cm−3
collimation of the OTR produced when the electron beam so nb,peak ≈ np,peak /2 putting the experiment just on the
emerges from the back of a polished aluminum foil. The boundary of the underdense regime, in terms of beam and
OTR is then transported either 30 cm to a conventional plasma peak densities, at the entrance to the plasma lens.
CCD camera or 7 m to a streak camera where the light is This is a lower bound however since the beam is substan-
imaged onto the streak slit using a single 280 mm focal tially denser compared to the edge of the gaussian plasma
length lens. column which it first encounters and, since the lens is thick,
The electron beam for the experiment is provided by the both analytic calculations and simulations show the beam
Fermilab NICADD Photoinjector Laboratory (FNPL) fa- core has compressed somewhat by the time it reaches the
cility [11]. The FNPL injector is a 15 MeV electron source center of the plasma lens. We can, therefore, unequivocally
consisting of a normal conducting L-band RF gun with a state that the beam focusing is governed by underdense dy-
Cs2 Te photo-cathode and a 9-cell superconducting post- namics at the threshold of the regime. When optimally fo-
accelerator cavity. After acceleration, the electron beam cused, a beam waist with xFWHM = 259±22 μm and yFWHM
is propagated at a tight focus through a 10 μm thick alu- = 423 ± 41 μm is achieved as shown in Fig. 2. The demag-
minum vacuum-isolation foil and refocused into the plasma nification factor of 6.29 in x and 3.65 in y means that the
lens experiment. Scattering in the aluminum foil signifi- transverse area of the focused beam is reduced by a fac-
cantly increases the beam emittance. The beam is trans- tor of 23. Note that the overall distribution of the plasma
ported 2.2 m from the vacuum foil to the center of the focused beam observed on OTR is not a single gaussian
but a superposition of the strongly focused core and unfo-
cused head and tail of beam, each of which remains roughly
(a) gaussian. By using the FWHM to characterize the focused
60 (c) X Axis
beam we automatically exclude the unfocused portions of
Projected Intensity (a.u.)

Plasma On

40
Plasma Off the beam (halo) which are visible as broad bases of the fo-
cused peaks in Fig. 2(c) and Fig. 2(d).
20 The effect of linear transverse focusing elements,
whether plasma or magnetic, on an electron beam traveling
0 in the z direction can be modeled (neglecting space charge
Plasma Off - 5 pulses -1.5 -1.0 -0.5 0 0.5 1.0 1.5
Transverse Position (mm) effects) by the beam envelope equation:
(b) 60 d2 σx 2x
(d) Y Axis + Kσ = , (1)
Projected Intensity (a.u.)

x
Plasma On
Plasma Off
dz 2 σx3
40
where σx is the rms beam size in x,  x is the geometric
20 beam emittance in x, and K is the focusing strength of the
lens. The equivalent equation for the other transverse di-
Plasma On - 1 pulse 0
-1.5 -1.0 -0.5 0 0.5 1.0 1.5
rection can be written by substituting y for x. For an un-
Transverse Position (mm) derdense plasma lens K = 2πr e np /γ, where re is the clas-
sical electron radius, and γ is the Lorentz factor [12]. De-
Figure 2: Images of the unfocused (a) and focused (b) elec- tailed predictions of the plasma lens focusing of the beam
tron beam OTR displayed with the same scaled to intensity core can be made by solving Eq. 1 using a function n p (z)
color map. In order to provide sufficient contrast, the unfo- that describes the shape of our thick plasma lens. The en-
cused image (a) is the integration of 5 beam pulses. The velope equation can also be used to describe the focusing
projected intensity of the focused and unfocused beams of lenses with aberrations. A general way to do this is to
(normalized to 1 pulse) in the x axis (c) and y axis (d) is define an effective emittance  eff = 20 + β0 0 δθ2 for use
also shown. in Eq. 1 which includes both the original emittance of the
03 Linear Colliders, Lepton Accelerators and New Acceleration Techniques A14 Advanced Concepts
1908 1-4244-0917-9/07/$25.00 2007
c IEEE
Proceedings of PAC07, Albuquerque, New Mexico, USA WEOAKI02

2000 360
1750
330
1500
300

xFWHM (µm)
xFWHM (µm)
1250

1000 270
750 240
500
210
250 Focused
Unfocused 180
-0.02 -0.01 0.01 0.02 0.03 0.04 -21 -20 -19 -18 -17 -16 -15 -14

Lens Center to OTR Distance (m) Beam Charge (nC)

Figure 4: Dependance of the focused beam spot size on the


Figure 3: Measurements of underdense plasma lens focus-
beam charge under otherwise similar conditions.
ing of a round beam in the x axis shown with solutions to
the envelope equation (solid lines and shaded region). The
shaded region indicate the shift in focal length caused by tion of a high brightness electron beam by a strong under-
the variation in plasma lens parameters. dense plasma lens and shown that, as predicted, this lens
has lower aberrations than overdense lenses even near the
boundary of the underdense regime (n b  np /2). We con-
beam 0 and the extra angular spread resulting from the jecture that the lens has a reasonably well formed ion col-
lens aberrations δθ [13]. The aberration
√ induced angular umn in this threshold regime due to the additive effect of
spread can be defined as δθ = ( β0 0 /f )(ΔK/K) where the electrostatic forces and the mutual repulsion between
β0 = (σ02 /0 ) is the beam beta-function at the lens en- the beam and plasma return currents. We are exploring the
trance, f is the lens focal length, and ΔK is the rms amount validity of this explanation through simulation. Operation
the lens strength deviates from linear expectation over the at the boundary of the underdense regime may be an at-
beam. Su et al. have shown that the minimum spherical tractive plasma lens scenario since it combines low aber-
aberration of an overdense plasma lens focusing a gaussian ration with minimal beam density. Lowering n b may be
beam is ΔK/K = 0.21 [2]. As stated previously, an under- one way to mitigate the problem of ion motion [14] for
dense plasma lens can theoretically be spherical aberration ILC class beams. A full analysis of this underdense plasma
free (ΔK/K = 0) in the ideal limit. lens experiment, including our time-resolved beam mea-
Fitting the predictions of the envelope equation to the surements, will be reported elsewhere [15].
focusing data obtained in the round beam case described
above, Fig. 3, reveals a great deal about the plasma lens REFERENCES
and its aberrations. The unfocused beam envelope is cal-
[1] P. Chen, Part. Accel. 20, 171–182 (1987).
culated using the initial conditions at the vacuum foil. The
plasma lens parameters vary somewhat as the mask is trans- [2] J. J. Su, et al., Phys. Rev. A 41, 3321 (1990).
lated, which leads to a slight variation in the lens focal [3] P. Chen, et al., Phys. Rev. Lett. 64, 1231 (1990).
length and, consequently, a broadening of the observed [4] K. Oide, Phys. Rev. Lett. 61, 1713–1715 (1988).
depth of focus as shown in Fig. 3. While the focal length [5] H. Nakanishi, et al., Phys. Rev. Lett. 66, 1870 (1991).
is determined by the known plasma lens parameters, the [6] G. Hairapetian, et al., Phys. Rev. Lett. 72, 2403 (1994).
size of the beam waist is determined by the parameter  eff [7] R. Govil, et al., Phys. Rev. Lett. 83, 3202 (1999).
which was obtained by a best fit to the data. For exam- [8] P. R. Bolton for the E150 Collaboration, Int. J. Mod. Phys.
ple, in the x direction the curve fit shown in Fig. 3 gives A 18, 2843 (2003).
eff,x,n = 110 mm-mrad, where we have followed conven- [9] J. S. T. Ng, et al., Phys. Rev. Lett. 87, 244801 (2001).
tion and used normalized emittance  n = βγ. A value
[10] M. C. Thompson, J. B. Rosenzweig, and G. Travish, Rev.
of x,n = 87 mm-mrad was measured by a quadrupole Sci. Instrum. 76, 013303 (2005).
scan downstream of the vacuum isolation foil. Using the
[11] J.-P. Carneiro, et al., “First Results of the Fermilab High-
above expression for  eff we can immediately calculate Brightness RF Photoinjector,” in Proc. PAC 1999, IEEE,
ΔKx /Kx = 0.076 ± 0.006 which is well below the over- 1999, p. 2027.
dense lens minimum of 0.21 and thus strongly indicative of [12] N. Barov, and J. B. Rosenzweig, Phys. Rev. E 49, 4407
underdense operation. The observed saturation of the fo- (1994).
cused beam spot size as the beam charge is increased to the
[13] J. B. Rosenzweig, et al., Part. Accel. 24, 11 (1988).
expected underdense threshold, Fig. 4, also supports the
[14] J. B. Rosenzweig, et al., Phys. Rev. Lett. 95, 195002 (2005).
case for underdense operation.
[15] M. C. Thompson, et al., Phys. Plasmas, (to be submitted).
In conclusion, we have measured extreme demagnifica-
03 Linear Colliders, Lepton Accelerators and New Acceleration Techniques A14 Advanced Concepts
1-4244-0917-9/07/$25.00 2007
c IEEE 1909